Download Life of a `Fact`: Understanding Infectious Disease Transmission

Life of a ‘Fact’: Understanding Infectious Disease Transmission
British Academy Conference: Enquiry, Evidence and Facts: An Interdisciplinary Conference
December 2007
Dr Erika Mattila
A simulated ‘fact’ uncovers indirect effects of vaccinations such as the reduction of bacterial circulation
in a population and the optimal vaccination coverage to reach herd immunity threshold. An agentbased simulation model is capable of revealing indirect, population level, effects on transmission:
effects from the type of vaccine used, the schedule followed, contact patterns and demographical
details. All these factors affect the disease transmission in a population. But how do these models
capture the finesses of disease transmission? In more general terms, the main focus of this study is to
explore the ways in which our understanding of the key questions in infectious disease studies evolved
over time and across disciplines. This will provide a way to reflect in a broader sense how ‘facts’ travel
across time and how useful the vehicle of mathematization is in that process.
Our story will show that transmission of disease is not covered by a single, solid account of the
process in which a susceptible person encounters a carrier and falls ill. There are various, pathogen,
population structure, age-specific and immunity related factors that need to be taken into account. On
the other hand, historical studies reveal how the understanding of transmission dynamics developed
into the current forms, which are oftentimes summarised in a generalised, pathogen-specific
transmission pattern that can be depicted in computational algorithms. Interestingly, the generalised
‘facts’ of transmission were to be tailored into a simulation model to produce sharpened facts of, for
example, precise transmission rates. The focus in this study is limited to the life of the transmission
‘fact’ and its capability to become adopted and adjusted in the mathematical forms presented through
modelling. The life history offers us a way to ask: Why do we know what we know of disease
transmission in epidemiology?
By examining the ways in which our understanding of disease transmission has developed through the
course of historical and current studies in epidemiology, we notice certain unsolved puzzles.
Transmission as a concept is itself related to a set of variables. These general ‘facts’ of transmission
have developed over time as our understanding has increased, but still, to examine the phenomena in
a detailed level, requires specification. Furthermore, we may ask whether we are able to identify
protofacts of transmission i.e. ideas that proceed the scientific facts and might even be hazy or
superfluous (e.g. Fleck 1979).
This study builds up a life history of the facts of disease transmission (cf. Daston 1999), which enables
us to understand the underlying currents of its birth, maturation, reaching of reproductive age and
passing away. This metaphor offers tools to depict the gradual, cyclic development and refinement of
‘facts’, instead of relying on the constructionist metaphors of scientists as carpenters in the business of
‘fact’ production. The analysis is based on historical resources on mathematical epidemiology, recent
textbooks and publications on mathematical modelling and its applications in infectious disease
studies, and ethnographic observations on modelling practices that led to the published Hib simulation
model. The analysis combines micro-sociological and micro-historical approaches in a narrative of a
life of a ‘fact’.
This study presents a life-history of disease transmission in the case of Haemophilus influenzae type b
bacteria, which is capable of causing severe diseases, such as bacterial meningitis in children. More
precisely, this story traces back the ways in which our understanding of disease transmission gradually
developed into the formulations captured in mathematical models used in current infectious disease
studies. In order to follow the general evolvement and refinement of the disease transmission on a
population level, the narrative will explore three distinct phases in the epidemiological research, all of
which refined and reiterated the concept of disease transmission in a population leading to the key
formulations used in current simulation models as force of infection or R0 (the basic reproductive rate).
The first phase will focus on the early development of mathematical epidemiology in the turn of the 19th
Century. Formulation of the germ theory of disease created a new framework to identify singular
causal factors behind infectious diseases. Koch’s postulates formed the generalised principles for the
conditions upon which an organism can be accepted as a cause of a particular disease. However, the
challenge remained how to address the disease transmission in a population. For mathematical
epidemiologists, the quest was to search for explanations for the “global patterns of disease in time,
space and population” (Fine 1979). This search resulted in understanding the cyclic patterns of
infections (e.g. Hamer 1906), refining the mathematical theory of epidemics that presented the
problem of which factors govern the “spread of contagious epidemics” (Kermack and McKendrick
1927), and in a misfortunate conception of infectiousness (by Brownlee in Fine 1979).
The second phase will follow the refinement of these theoretical and experimental observations of
disease transmission and explore how they become simplified and sharpened into probabilistic
infectious disease models. A special focus is on the disease patterns which classify individuals into
susceptible, infected and recovered groups (or different variants of these). These patterns are usually
expressed as SIS, SIR or SEIR (susceptible, infectious, infected, and recovered) depending on the
bacterial agent in question. In our story the pattern is SIS since Hib does not convert to permanent
immunity. Interestingly though, these patterns form a generalised body of knowledge that facilitates the
parameterisation of infectious disease models. However, in the early phases of epidemiological
modelling, this classification led to deterministic compartmental models of disease transmission. These
models are insensitive to the impact of chance in small populations and oftentimes ignore agent-based
dynamics. So, the second phase of the exploration will clarify the difference between deterministic and
probabilistic models and discuss the ways of understanding disease transmission in a population.
The third phase ends our story with a structured simulation model, which is capable of capturing agentbased dynamics, addressing mixing patterns and including vaccination effects on the circulation and
transmission of Hib. This fine-grained, population simulation model is hence capable of telling us the
full notion of Hib transmission, immunity and disease (Auranen et. al. 2004). Our story will take a
detailed look at the transmission dynamics captured by the model and discuss the ways in which
transmission is represented, simplified and quantified in the model assumptions and parameters.
Through these three phases that form the backbone of our narrative, we will learn how the
understanding of disease transmission evolved from the discovery of germs as singular cause of
disease into the sophisticated ways of simulating disease transmission in population. The main
findings discuss what kinds of translations ‘facts’, originating in epidemiological studies, face in order to
be simplified, refined and ‘mathematized’ into models.
References:
Auranen et. al. 2004: Modelling transmission, immunity and disease of Haemophilus influenzae type b
in a structured population. Epidemiology and Infection (2004), 132, 947-957.
Daston (eds.) 1999: Biographies of Scientific Objects. University of Chicago Press.
Fine 1979: John Brownlee and the Measurement of Infectiousness: An Historical Study in Epidemic
Theory. Journal of the Royal Statistical Society. Series A (General), Vol 142:3, 347-362.
Fleck 1979/1935: Genesis and Development of a Scientific Fact. University of Chicago Press.
Hamer 1906: Epidemic Disease in England. The Lancet i, 733-9.
Kermack & McKendrick 1927: A contribution to the mathematical theory of epidemics. Proceedings
from the Royal Society A115, 700-21.